Brain, Vol. 125, No. 9, 2067-2080,
September 2002
© 2002 Guarantors of Brain
Can segregation within the semantic system account for category-specific deficits?
1 The Wellcome Department of Cognitive Neurology, Institute of Neurology, London and 2 Behavioural Brain Sciences, School of Psychology, University of Birmingham, Birmingham, UK
Correspondence to: Dr Cathy Price, Wellcome Department of Cognitive Neurology, 12 Queen Square, London WC1N 3BG, UK E-mail: c.price{at}fil.ion.ucl.ac.uk
Received November 1, 2001. Revised March 4, 2002. Accepted March 6, 2002.
 |
Summary |
|---|
 Top Summary IntroductionMethodsResultsDiscussionAppendixReferences |
|---|
Functional neuroimaging was used to investigate the extent to which category-specific semantic deficits in patients can be accounted for in terms of the demands placed on neural systems underlying different types of semantic knowledge. Unlike previous functional imaging studies of category specificity, we used a factorial design that crossed category (tools and fruits) with tasks requiring retrieval of either action or perceptual (real life size) knowledge. The presentation of tools relative to fruit increased activation in the same left posterior middle temporal area that was linked to the retrieval of action knowledge in general (for fruit as well as tools). However, we found no correlation between activation evoked by fruit and the size retrieval task. The left medial anterior temporal cortex was the only region to be activated for fruit relative to tools. We argue that the sensory–functional theory of category-specific effects is insufficient to account for the current neuroimaging literature. However, the data do support a more refined version of the theory: tools, relative to fruit, are more strongly linked to manipulative/motor knowledge and, for some tasks, fruit may be more reliant on integrating multiple semantic features.
Keywords: category specificity; semantics; functional imaging
Abbreviations: LPMT= left posterior middle temporal cortex; RT = reaction time; SMA = supplementary motor area; SPM = statistical parametric mapping; STS = superior temporal sulcus
|
Introduction |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionAppendixReferences |
|---|
The observation that patients with focal brain damage can differ in their ability to identify living and non-living entities was suggested but not quantified by several early investigators (Nielsen, 1946
; Hécaen and Ajuriaguerra, 1956; Yamadori and Albert, 1973). After a more comprehensive study of this phenomenon, Warrington and Shallice (1984) developed a theoretical framework that accounted for deficits with living things (e.g. animals and fruits) in terms of impaired perceptual knowledge, whereas deficits with non-living things (e.g. tools and vehicles) were explained in terms of impaired functional knowledge. This perceptual–functional theory was based on observations that the patients who exhibited deficits with living entities were able to identify body parts which are functional components of living things but were unable to identify different fabrics and gem stones, i.e. non-living items which are discriminated on the basis of perceptual knowledge. In 1987, Warrington and McCarthy extended the theory to account for the behaviour of another patient (Y.O.T.) who had a generalized deficit for non-living over living items, but was more impaired at identifying smaller ‘manipulable’ items such as office tools and kitchen utensils than larger man-made objects such as vehicles. In light of these findings, Warrington and McCarthy (1987) proposed a revised model, suggesting that object-specific knowledge is acquired according to the relative salience of its sensory–motor features and is thus weighted towards the dominant ‘channel’ of experience; sensory-, motor- or functional-associative (see also Humphreys and Forde, 2001). In the case of tools, and body parts, the dominant channel of experience involves motor (i.e. manipulative action) and functional (i.e. knowing the context of usage) information. A plausible biological embodiment of this model can be envisaged in terms of the different ‘channels’ manifested in the human brain as regions of cortex specialized to receive different modalities of sensory input, or initiate motor output from different effectors.
Increasingly, there is evidence that motor and functional knowledge can be doubly dissociated. For instance, Buxbaum et al. (2000) report two apraxic patients who could match items on the basis of their function (a piano and a record player), but not on the basis of how they are manipulated (a piano and a typewriter). In contrast, Sirigu et al. (1991) report a patient, F.B., with associative agnosia, who could retrieve object-specific actions but not function knowledge. If motor/action knowledge is crucial to the categorization and identification of tools, one would expect that patients like F.B. (with preserved motor knowledge) would still be able to identify and name tools. Indeed, Sirigu et al. (1991) noticed that when F.B. gestured the use of an object, he was much more successful at describing it verbally, and this observation has been corroborated by descriptions of naming behaviour in disorders such as optic aphasia (Riddoch and Humphreys, 1987; Magnié et al., 1999). Conversely, one would expect that impaired motor knowledge (e.g. in patients with apraxia) should be associated with difficulties identifying tools. Some suggestive evidence is provided by Buxbaum and Saffran (1998), who found that a group of patients with moderate apraxia also had difficulty recognizing tools, but no difficulties recognizing pictures of animals. Further investigations are required to confirm these findings.
Although there is considerable evidence to support the association between tool processing and manipulative knowledge, the evidence supporting the link between perceptual knowledge and the category of natural kinds is less convincing. The primary problem with this position is that not all patients with deficits identifying natural kinds have a corresponding loss of sensory or perceptual knowledge (Forde et al., 1997; Caramazza and Shelton, 1998; Lambon Ralph et al., 1998; Moss et al., 1998). Indeed, category-specific impairments for living things could arise potentially at multiple levels of processing, for instance due to different demands on pre-semantic perceptual processing (see Humphreys and Forde, 2001). In the study presented here, we test whether the perceptual/functional and sensory/motor theories might still be a plausible biological account of some category-specific effects.
Neuroimaging evidence
Early neuroimaging experiments investigating the effects of object category offered some support for the sensory–functional/motor theory proposed by Warrington and Shallice (1984) and Warrington and McCarthy (1987). Activation for animals relative to tools was greater in visual association areas (Perani et al., 1995; Martin et al., 1996), and activation for tools relative to animals was greater in a left posterior middle temporal (LPMT) region (Martin et al., 1996; Mummery et al., 1996) which has also been associated with retrieving action relative to colour words (Martin et al., 1995). Taken together, the studies by Martin et al. (1995, 1996) suggested that tools increase the demands on an area associated with ‘action knowledge’. Several subsequent studies have replicated the association of LPMT with tools (Damasio et al., 1996; Mummery et al., 1996, 1998; Cappa et al. 1998; Moore and Price, 1999; Perani et al., 1999). However, other effects of category and type of semantic knowledge have been inconsistent (Vandenberghe et al., 1996; Cappa et al., 1998; Mummery et al., 1998; see Moore and Price, 1999). For instance, the visual association areas, linked to the category of animals, are not seen when the complexity of visual input is controlled (Moore and Price, 1999) or the stimuli are written words (Cappa et al., 1998; Mummery et al., 1998), suggesting that these regions are concerned with early visual processing of pictures rather than amodal semantic processing.
Ideally, in order to establish that category effects emerge from the differential demands placed on different channels of semantic knowledge, a factorial design is required that crosses category with type of knowledge. It can then be established whether anatomical segregation for different categories corresponds to the anatomical segregation for types of knowledge. Three previous studies have already taken this approach (Cappa et al., 1998; Mummery et al., 1998; Thompson-Schill et al., 1999). Tasks probing perceptual knowledge involved retrieval of object colour (Mummery et al., 1998) or object form (Cappa et al., 1998; Thompson-Schill et al., 1999) but only the study of Thompson-Schill et al. (1999) reported an association between the perceptual task and category. They found that their region of interest (the left anterior fusiform—previously linked to visual knowledge by D’Esposito et al., 1997) was activated by living and non-living items when the questions were visual–perceptual but only by living items when the questions were not perceptual. They interpreted this finding as evidence that the semantic representation of living items is biased toward perceptual knowledge irrespective of the task requirements. However, due to design constraints, the critical difference between living and non-living items in the functional-associative condition could not be reported; and the category by task interaction was exceedingly small.
The factorially designed experiments conducted by Cappa et al. (1998), Mummery et al. (1998) and Thompson-Schill et al. (1999) also failed to find any association between non-living items and functional-associative knowledge. Mummery et al. (1998) used associations based on typical location (‘Is PAPER found in the same place as a RULER or a RAKE?’); Cappa et al. (1998) used functional questions for tools (‘Is it used for food preparation?’) and location questions for animals (‘Is it from Italy?’); and Thompson-Schill et al. (1999) used more specific questions such as ‘Can headphones play stereo music?’. These consistent null findings do not favour an equivalence between non-living effects and the demands on functional-associative knowledge, but the possible correspondence between tool identification and action/motor knowledge was not investigated.
Rationale for study design
In contrast to the previous neuroimaging studies that focused on functional-associative knowledge, our design aimed to identify the cortical regions that mediate action retrieval. LPMT is one area that has been identified in action-word generation (Martin et al., 1995), action recognition and action observation tasks (see Grézes and Decety 2001), suggesting that it plays a role in praxis and action planning. The studies of Martin et al. (1995, 1996) indicated that similar posterior temporal regions are involved in action and tool identification (see above), but a between-study statistical validation has never been reported. Furthermore, in these studies (Martin et al., 1995, 1996), activation for neither action nor category reached a level of significance that was corrected for the number of comparisons made. The application of this threshold has now become standard practice in neuroimaging experiments because of the potentially high false-positive rate. As a result, there are doubts concerning the validity of most of the category-specific effects reported in the neuroimaging literature (for a detailed review see Devlin et al., 2002a).
Our study was designed to explore the extent to which the neural substrates for action/motor knowledge and tool processing are equivalent. Retrieval of action knowledge was compared with retrieval of perceptual (size) knowledge, and the stimuli were pictures and written words depicting either tools or fruit. We chose to use constrained decision tasks (e.g. ‘Do you use a twisting motion to manipulate this tool?’; ‘Can you peel this fruit by hand?’; ‘Is the tool longer than a paintbrush?’; and ‘Is the fruit larger than a kiwi?’) rather than word generation tasks (Martin et al., 1995) that may be confounded with strategic, phonological and articulation differences. For example, the action and colour generation tasks used by Martin et al. (1995) were not matched for availability of response, as most non-living stimuli (e.g. pencils) can appear in many different colours but only have one stereotypical action. Finally, by presenting stimuli as both pictures and words, we could look for effects of stimuli or task that were independent or specific to presentation modality (words or pictures).
Predictions
On the basis of the studies by Martin et al. (1995, 1996), we predicted that tools (relative to fruit) and retrieval of action (relative to perceptual) knowledge independently would enhance activation in the LPMT. Although other investigators have proposed that the left premotor cortex is involved in retrieving action knowledge (Martin et al. 1995; Grafton et al., 1997) and in tool processing (Martin et al., 1996; Grabowski et al., 1998; Chao and Martin, 2000), differences between tools and fruits have not been observed (see Devlin et al., 2002b). Furthermore, a recent meta-analysis of action-related activation in PET experiments (Grézes and Decety, 2001) has verified that only the activation reported by Grafton et al. (1997) can be allocated to the dorsal precentral gyrus (the hand area; Fink et al., 1997). Other reported activations (Martin et al., 1995; Grabowski et al., 1998) verge on the ventral precentral mouth area, and may be due to silent verbalization rather than action ideation. Thus, it is not clear whether the premotor area will be activated differentially by either (i) action relative to perceptual tasks or (ii) tools relative to fruit.
In addition, we anticipated that fruit (relative to tools) might enhance activation in the anterior temporal lobes. Damage to these regions has been associated with category-specific deficits for living things (see Gainotti et al., 1995). Three previous functional neuroimaging studies (Mummery et al., 1996; Moore and Price, 1999; Devlin et al., 2002b) have also found these regions to be more active for living than non-living things.
|
Methods |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionAppendixReferences |
|---|
Subjects
Twelve male subjects (age range 21–45 years) took part in this experiment; they were all right handed native English speakers, who were healthy, on no medication and free from any history of neurological illness. Subjects’ consent was obtained according to the Declaration of Helsinki, and the study was approved by The National Hospital for Neurology and Neurosurgery and Institute of Neurology Joint Research Ethics Committee and the Administration of Radioactive Substances Advisory Committee (UK) (ARSAC).
Data acquisition
Each subject underwent 12 PET relative perfusion scans over a 2-h period. Scans were obtained using a SIEMENS/CPS ECAT EXACT HR+ (model 962) PET scanner (Siemens/CTI, Knoxville, TN, USA) with collimating septa retracted. Participants received a 20 s intravenous bolus of H215O at a concentration of 55 MBq/ml and at a flow rate of 10 ml/min through a forearm cannula. For each subject, a T1-weighted structural MRI was obtained with a 2 T Magnetom VISION scanner (Siemens, Erlangen, Germany).
Conditions and tasks
There were 12 conditions in a 2 x 2 x 3 factorial design. The first factor was stimulus type: pictures of objects or written object names. The second factor was stimulus category: natural kinds (fruits or vegetables) or artefacts (tools). The third factor was task, which required subjects to make decisions on (i) retrieved action; (ii) retrieved size; or (iii) the screen size of the stimulus. In the action retrieval task, subjects were asked if they might use a twisting motion when manipulating the presented tools, or if the fruits and vegetables could be peeled by hand (e.g. remove the skin of a banana). They indicated their choice (‘yes’ or ‘no’) by pressing a key with the index or middle finger of the right hand (e.g. left key for yes and right for no). The size retrieval task required imagination of the dimensions of the object in the real world. For example, ‘larger than a kiwi?’ was one of the questions applied to a set of fruit stimuli, and ‘longer than a fork?’ was one of the questions applied to tools. A full list of questions and the sets of objects to which they were applied is provided in the Appendix. The screen size decision task was used as a baseline for the semantic retrieval tasks. Subjects decided whether the picture or word presented was wider or narrower than the length of a horizontal line drawn beneath it (the line was present during all conditions, but subjects attended to it only during screen size judgements). Subjects were trained to make these judgements prior to scanning. Reaction times and accuracy of performance were recorded.
Stimuli
The stimuli were selected from a set of 48 line drawings of objects from the Snodgrass and Vanderwart (1980) picture set. Half of the drawings depicted tools, and the other half fruit and vegetables. Each object was presented to one group of six subjects as a word, and to the other group of six subjects as a line drawing. The stimuli were matched for frequency, word length and number of syllables (see Appendix).
Procedure
There were a total of 12 conditions and 12 scans per subject (i.e. a new condition for each scan). Each of the four stimulus sets (names of fruit/vegetables, names of tools, pictures of fruit/vegetables and pictures of tools) was presented three times during the experiment, once for each task. This was to ensure that between-task differences were not due to differences between stimuli. The order of replication was counter-balanced between subjects. Within subjects, the effects of priming between conditions were reduced by requiring subjects to name or read the names of all 48 stimuli immediately prior to scanning. Each trial was 1 s in duration, followed by a 3.5 s interstimulus interval. The total stimulus presentation period was 1 min, starting 10 s before data acquisition, and 12 stimuli were shown per condition (four cued a ‘yes’ response and eight cued a ‘no’ response). Before each scan, subjects were presented with written instructions which detailed the category (tools or fruit) and modality (words or pictures) of stimuli to be presented and the task to be performed. ‘Yes’ and ‘no’ responses were randomized within scans, and the modality and category of the stimuli (as a word or picture, living or non-living entity) were fully counter-balanced across subjects. The order of conditions was also balanced over subjects. Stimulus presentation was via a Macintosh computer screen at a distance of 40 cm from the subject. The stimuli subtended an average visual angle of 6.1° for words and 8.4° for pictures.
Data analysis
The data were analysed with statistical parametric mapping (SPM; using SPM99 software from the Wellcome Department of Cognitive Neurology, London, UK; http//www.fil.ion.ucl.ac.uk/spm) implemented in Matlab (Mathworks Inc. Sherborn, MA, USA) using standardized procedures (Friston et al., 1995a, b) including realignment, spatial normalization (voxel size = 2 mm3), smoothing and ANCOVA (analysis of covariance). The smoothing kernel was a 3D Gaussian filter of 16 mm. Condition and subject effects were estimated according to the general linear model at each voxel. To test hypotheses about regionally specific condition effects, these estimates were compared using linear compounds or contrasts. The resulting set of voxel values for each contrast is an SPM of the t statistic, which was then converted to a Z statistic. Unless otherwise stated, SPMs were thresholded at P < 0.05 corrected for multiple comparisons and the threshold for masks was set at P < 0.001 uncorrected.
The linear contrasts
Common effects of semantic retrieval
The regional activations common to both action retrieval and size retrieval were identified by comparing the sum of the semantic retrieval conditions (action and size over category and stimulus modality) with the screen size control conditions. To ensure that both action and size retrieval decisions were contributing to the effect, inclusive masking (SPM99) was applied to the resulting statistical map. The masks specified were (i) action retrieval minus the screen size control and (ii) size retrieval minus the screen size control. Inclusive masking includes only those voxels that are activated in all the contrasts specified.
This same procedure was repeated including (i) word conditions only and (ii) picture conditions only. Differences between words and pictures were then explored by testing for the interaction between stimulus modality (words versus pictures) and task (action and size retrieval versus screen size decision).
Effects specific to action or size retrieval
Regional activations specific to action retrieval were identified by comparing this with size retrieval. We ensured that regions of activation revealed in this contrast were also significantly greater than in the control condition by inclusively masking action minus size retrieval with action minus the screen size control. Likewise, areas specific to size retrieval were identified by masking size retrieval minus action retrieval with size retrieval minus the screen size control.
We also tested for interactions between task (action versus size retrieval) with (i) stimulus type (words versus pictures) and (ii) stimulus category (tools versus fruit).
Category effects
The effect of tools relative to fruit (and vice versa) was computed for each task separately. This allowed us to look for task-specific effects. Conjunction analyses (see Price and Friston, 1997) were then used to test for common effects (i) over all tasks and (ii) over semantic tasks only.
Differential effects of category were identified by testing for the interaction of category with stimulus type and task.
|
Results |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionAppendixReferences |
|---|
Behavioural data
The mean error rates and mean reaction times (RTs) over all subjects are shown in Table 1 after exclusion of scores >2.5 SD from the mean.
|
RTs were analysed with a within-subjects ANOVA (analysis of variance) to test for significant differences between the conditions. The only significant effect pertained to an interaction between task and modality F(2,20) = 4.3; P < 0.0285 with RTs faster for the baseline task on words than pictures.
Imaging data
Common effects of semantic retrieval (see Table 2 and Fig. 1)
Action and size retrieval, relative to the screen size control, activated (i) the left inferior and middle temporal gyri (BA 20 and 21); (ii) the left inferior frontal cortex within BA 47 (below the diagonal limb of the lateral fissure), BA 45 (pars triangularis) and BA 44 (pars opercularis); and (iii) the left fronto-marginal gyrus (BA 10/12) (see green regions in Fig. 1). There were no effects of semantic retrieval that were specific to words or pictures.
|
Effects specific to action or size retrieval (see Tables 3 and 4)
Activations specific to action retrieval (see red regions in Fig. 1) were localized to our region of interest in the LPMT (BA 37/21) and the right posterior-medial cerebellum (HVII A; Schmahmann et al., 1999
). There was also a trend for increased activation in the left ventral inferior frontal cortex (P < 0.001 uncorrected; see orange area in Fig. 1). However, there was no effect of action relative to size retrieval in the premotor cortex, even when we lowered the threshold to P < 0.05 uncorrected.
Activation specific for size retrieval was observed only on the medial surface of the right superior frontal gyrus [pre-supplementary motor association cortex: designated as such because it is located in front of the VAC (vertical anterior commissural) line; Picard and Strick, 1996], see Fig. 2A and B.
|
There were no significant interactions between task (action or size) and stimulus type or category.
Category effects
The only effects that survived a corrected level of significance were for tools relative to fruits during semantic tasks (see Table 5). These were located in (i) the right medial premotor cortex (supplementary motor area; SMA) behind the VAC line in accordance with Picard and Strick (1996) (see Fig. 3A and B); and (ii) LPMT where the effect was just below a corrected level of significance (Z = 4.6; P < 0.06 corrected; see Fig. 2B and C). The respective Z scores for the interaction of category (tools–fruit) and task (semantic–screen size) in these areas were 3.6 in the SMA and 2.33 in the LPMT.
|
|
The effect of tools relative to fruit in LPMT overlapped with the effect of action relative to size retrieval. More specifically, the effect of tools relative to fruit was present for both action (x = –58, y = –64, z = 4; Z = 3.7) and size (–54, –70, –4; Z = 3.7) retrieval tasks; and the effect of action relative to size retrieval was present for both tools (–58, –60, 4; Z = 4.4) and fruit (–52, –70, –6; Z = 4.1). Figure 2 illustrates the overlap of task and category in LPMT.
We then adopted an uncorrected significance threshold (P < 0.001) to see if there was any activation for fruits relative to tools in the anterior temporal cortices, where lesions are associated with category-specific deficits for natural kinds (Gainotti et al., 2000). We found greater activation for fruits than tools in the left medial anterior temporal cortex (x = –22, y = 2, z = –24; Z = 3.02; P < 0.001 uncorrected) irrespective of task. There were no significant effects in the right anterior temporal cortex even when the threshold was lowered to P < 0.05 uncorrected. Furthermore, there was no trend towards increased anterior temporal activation for the size retrieval task relative to either the action or control tasks (P > 0.05 uncorrected).
Finally, there were no significant effects of tools relative to fruit in the left premotor cortex at a corrected or uncorrected (P < 0.001) level. However, when the threshold was lowered to P < 0.05 uncorrected, a trend was observed (x = –42, y = 4, z = 12; Z = 2.1; P < 0.05 uncorrected).
|
Discussion |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionAppendixReferences |
|---|
Our aim was to establish whether the regions involved in processing tools and fruit correspond to those involved in the retrieval of action and perceptual (size) knowledge, respectively. Support for this claim for tools was indicated by greater activation in LPMT for (i) retrieving action relative to size for both tools and fruit; and (ii) tools relative to fruit for both action and size retrieval tasks. However, there was no overlap in the activations observed for size relative to action retrieval, and fruit relative to tools. Fruit alone activated the left medial anterior temporal cortex (as predicted), and the only effect of size retrieval was in the pre-SMA.
Unpredicted, but highly significant activation was also observed in the right cerebellum for action relative to size retrieval, and in the SMA for tools relative to fruit. We discuss these results in terms of task-specific and category-specific effects. By reference to the previous literature, we distinguish activations that are specific to retrieval of action or size from those that reflect differential demands on more general retrieval strategies. Finally, we discuss the significance of these results for the cognitive and anatomical models of semantic processing that attempt to explain category-specific semantic deficits.
Task effects in LPMT
The results confirmed our prediction that LPMT is involved in action retrieval; there was no detectable activation in LPMT for size retrieval, even at a very liberal threshold (P < 0.05 uncorrected). Although focal damage to LPMT has not been widely associated with tool-specific deficits in patients (except by Tranel et al., 1997), other neuroimaging data consistently suggest that LPMT has a role in the imagination and planning of actions.
LPMT is activated by retrieval of action words (Martin et al., 1995; Fiez et al., 1996; Warburton et al., 1996) but not perceptual (e.g. colour words; Martin et al., 1995) or associative (Vandenberghe et al., 1996) knowledge. Nevertheless, LPMT responses are not limited to the retrieval of action; LPMT is also activated when actions are observed, e.g. when subjects view moving images of hand gestures relative to static images of hands (Decety et al., 1994, 1997; Grezes et al., 1998, 1999), or eye and mouth movement relative to non-biological motion (Puce et al., 1998). LPMT also responds when actions are implied, e.g. when static images of hand gestures are compared with natural scenes (Peigneux et al., 2000), and static images of objects falling and athletes running are compared with similar images in which no motion is implied (Kourtzi and Kanwisher, 2000; Senior et al., 2000). Although pictures of animals can also imply motion, LPMT typically is more active for pictures of tools (Damasio et al., 1996; Martin et al., 1996; Moore and Price, 1999); therefore, it is not simply linked to biological motion. Neither is the response in LPMT limited to familiar actions that have ‘semantic meaning’, as several studies report equivalent activation in this area for meaningless and meaningful gestures (Decety et al., 1997; Grezes et al., 1998; Peigneux et al., 2000). Likewise, we have shown LPMT activation when action decisions are made to novel objects (Phillips et al., 2002).
Cumulatively, the evidence suggests that LPMT is involved when stimuli show or imply human movement. Could this area be part of a neural system for praxis and action ideation? Rizzolatti et al. (1996) suggest that human LPMT is homologous to the lower bank of the superior temporal sulcus (STS) in monkeys. Single neurones in the monkey STS fire selectively in response to behaviourally relevant motion such as hand actions, and their firing rates are also modified by gaze direction (Perrett et al., 1989; Oram and Perrett, 1996). In humans, LPMT and the lower bank of the STS may be functionally continuous as they share Brodmann’s area 21. Indeed, our LPMT activation for action retrieval and tools extended into the lower bank of the STS. Alternatively, LPMT may be involved in a more basic level of action analysis and then transfers this information to STS where more complex analysis of intention has been shown to occur in humans (e.g. Castelli et al., 2000). White matter connections between STS and LPMT (Seltzer and Pandya, 1978, 1994) might facilitate such an interaction. Another possibility is that LPMT is homologous to the motion detection region (V5/MT) which also responds to implied or illusory motion (Zeki et al., 1993; Kourtzi and Kanwisher, 2000; Senior et al., 2000). The average coordinates of V5 (x = –41 ± 5.6 mm, y = –70 ± 6.0 mm, z = +2 ± 5.3 mm; Watson et al., 1993) are 8 mm posterior to the LPMT area we associate with action retrieval, and 20 mm posterior to the area that Martin et al. (1995) associated with action retrieval (x = –50 mm, y = –50 mm, z = +4 mm). As these are neighbouring regions, further investigation is needed to determine whether their neural populations are indeed functionally separable.
In summary, the evidence presented suggests that actual or implied human actions increase the response in LPMT: e.g. when body parts (e.g. hands) or manipulable objects (e.g. tools) are involved in a gesture; when a gestural/manipulative movement is implied (e.g. static images of hands gesturing, athletes running, objects falling); or when an action is verbally planned (e.g. action decisions, action-word generation). LPMT does not appear to be affected by the familiarity of the movement. Nevertheless, when the stimuli are meaningful, it is usually activated in conjunction with the neural systems associated with semantic retrieval.
Task effects in right cerebellum and pre-SMA
While responses in LPMT are strongly associated with action-related stimuli and tasks, the previous literature does not support such a role for the right cerebellum. We found greater activation in the right cerebellum for action relative to size retrieval, and no effect of size retrieval relative to screen size judgements. However, in contrast to the present study, Vandenberghe et al. (1996) report that size retrieval relative to screen size judgements did activate the right cerebellum (x = 38, y = –76, z = –48); therefore, this region is not specific for action retrieval. The inconsistent effects for size retrieval in our study and that by Vandenberghe et al. (1996) might be explained by the hypothesis proposed by Desmond and Fiez (1998), who have argued that right cerebellar activation during verb generation tasks (Petersen et al., 1988, 1989; Klein et al., 1995; Martin et al., 1995; Fiez et al., 1996) relates to ‘semantic [retrieval] difficulty’. However, in our study, differences in retrieval difficulty for the action and size tasks were not reflected by differences in reaction times (see Table 1). Thus, the activation in the cerebellum may be related to a cognitive component of task performance other than action retrieval or semantic difficulty.
Likewise, the response in the pre-SMA (more active for size relative to action retrieval) should not be interpreted as specific to the retrieval of perceptual knowledge. Pre-SMA activation has been reported during a variety of tasks in which information must be held on-line in preparation to execute a response, such as recalling spatial locations and recognizing faces (for a review see Petit et al., 1998). It is also activated during the retrieval and performance of complex motor sequences in both monkey and man (Tanji and Shima, 1994; Sakai et al., 1999).
It is noteworthy that both of these regions, in association with the inferior frontal cortex, have been identified as parts of a neural system involved in retrieving items from verbal/semantic working memory (e.g. Fiez et al., 1996; Gabrieli et al., 1997). However, to our knowledge, a double dissociation of their respective functions has never been demonstrated before. Based on the evidence above, we attribute the activations in the right cerebellum (associated with action retrieval) and the pre-SMA (associated with size retrieval) to cognitive processes that are not specific for retrieving action or size knowledge, but which may be related to working memory.
Thus, one explanation for the task-specific activations in the pre SMA and the right cerebellum might be differential demands upon working memory rather than activation for action and size retrieval. This interpretation is borne out when considering the requirements of each task beyond retrieving action or size knowledge. For example, in the size retrieval condition, each scan would begin with a different question, e.g. ‘Is the object longer than a paintbrush?’. The subject was required to hold this information on-line, in working memory, for the duration of the block, in order to compare the length of a stereotypical paintbrush with that of each stimulus presented subsequently. In contrast, action retrieval did not tax working memory to the same extent because the questions presented were always the same (i.e. ‘Can this tool be twisted during use?’/‘Can this fruit be peeled by hand?’). Further research is required to investigate whether the pre-SMA and right cerebellar regions are affected differentially by the demands on working memory. With respect to the present study, however, we can conclude that activation in the pre-SMA and the right cerebellum is not specific to the type of information retrieved.
In summary, both action and size retrieval activated a generalized semantic retrieval system (see Fig. 1) that is consistent with the pattern of activation reported by Vandenberghe et al. (1996), Mummery et al. (1996, 1998) and Devlin et al. (2000). In addition, (i) retrieving action knowledge activated a region in LPMT; (ii) effects attributed to differences in task strategy were found in the right cerebellum (for action retrieval) and the pre-SMA (for size retrieval); and (iii) no areas were revealed that could be linked specifically to perceptual knowledge. This null result for living entities and perceptual knowledge might also be attributable to the perceptual task we employed, i.e. the real size dimension might be less relevant than shape or colour for distinguishing between living/natural entities.
Category effects: tools
In accordance with the account of category specificity that stresses differences between sensory and motor (action) knowledge for living and non-living things, respectively (Magnie and McCarthy, 1987), we found an additive effect of tools and action in LPMT. The semantic retrieval tasks with tools also activated the right rostral-medial premotor cortex (SMA) where there was an interaction between task (semantic > non-semantic) and stimulus category (tools > fruit) at an uncorrected P < 0.001 level. The SMA is thought to be involved in higher motor function (Picard and Strick, 1996) and is directly connected with the primary motor cortex (Dum and Strick, 1991), allowing direct access to the motor effectors. Activation in the rostral sector of the SMA, located closer to the VAC line, is reported consistently during the mental rehearsal of movement (Stephan et al., 1995; Fink et al., 1997; Christensen et al., 2000). Thus, co-activation of the rostral SMA and LPMT for the semantic processing of tools more than fruit suggests that semantic processing of tools implicitly prompts imagery of motor action. The SMA may be involved in action preparation, whereas LPMT appears to be involved in the visual analysis of the stimulus for its action propensity (see above). However, no previous functional imaging study has linked the SMA to tool-specific activation, suggesting that activation may be specific to the type of semantic tasks we employed (action and real life size decision). Lesion data offer further support because focal lesions to the SMA can result in apraxia. Watson et al. (1986) describe two such patients who have a bilateral apraxia for transitive movements (i.e. involving tool use, such as hammering), but not intransitive movements (i.e. symbolic gestures such as hitchhiking that do not involve tools or objects). Although damage to the SMA has not been linked specifically to category deficits, patients with deficits related to non-living entities are often reported to have sustained fronto-parietal lesions (Gainotti et al., 2000) which may include the SMA or the connections to it.
Finally, there was negligible activation for tools in the left premotor cortex (Z = 2.1; P < 0.05 uncorrected) compared with the strong effects observed in LPMT (Z = 4.6; P < 0.06 corrected; P < 0.0000 uncorrected). Similarly, the left premotor effects reported when observing (Chao and Martin, 2000) or naming (Martin et al., 1996) tools relative to animals were not highly significant. Thus, category-specific effects in this area appear to be small and may depend on the task or stimuli. For example, Devlin et al. (2002b) have suggested that tools may activate the left premotor cortex relative to animals but not relative to fruit, which are also manipulable. Alternatively, in our study, subjects made a manual motor response (key press) in all conditions, thereby equating manual response, whereas when the response is non-manual (e.g. during viewing or naming), tools may elicit implicit manual responses that are incidental to the task requirements.
Category effects: fruit
The only area to respond to fruit more than tools was the left medial anterior temporal lobe where damage can sometimes result in category-specific impairments (Gainotti et al., 2000). Activation was significant at P < 0.001 uncorrected in a predicted region. Nevertheless, this effect is not as robust as the LPMT activation we observed for tools. Why are fruit-related effects in the anterior temporal cortices relatively elusive? Our explanation is that responses within the medial anterior temporal cortex are not specific to processing natural kinds of objects but depend on the demands placed on differentiating or integrating multiple semantic features. Functional imaging evidence that anterior temporal activation is involved in ‘differentiating’ semantic features comes from a recent study by Devlin et al. (2002b). In this study, the left medial anterior temporal region (x = –28, y = 2, z = –26) was found to be activated by tools as well as fruit during semantic categorization tasks which involved deciding whether the last word of a sequence belonged to the same category as the preceding words. For example, when given the word sequence ‘knife, spoon, fork, MARBLES’, the response was ‘no’ because ‘MARBLES’ are not a member of the cutlery category. The evidence that anterior temporal activation is involved in ‘integrating’ semantic features comes from studies showing that left anterior temporal activation increases when sentences are compared with unrelated word lists (Bottini et al., 1994) and when stories are compared with unrelated sentences (Fletcher et al., 1995).
With respect to differential effects of category, the anterior temporal cortex may be more active for identifying natural relative to man-made items (Mummery et al., 1996; Moore and Price, 1999; Devlin et al., 2002b; this study) because identifying natural objects, such as fruit, is more reliant on integrating multiple features, whereas man-made objects, such as tools, can be identified on the basis of individual distinctive features (Devlin et al., 1998; McRae and Cree, 2002). For instance, identification of fruit usually relies on more than one property. Knowing that a fruit is yellow or long, or sweet or sour, is not sufficient, but knowing that it is yellow and sour (lemon) or yellow and long (banana) is sufficient. In contrast, knowing a single property about a tool (e.g.‘drives nails’) can often be sufficient to identify the object (HAMMER).
In summary, we are arguing that (i) the medial anterior temporal cortex appears to be involved in differentiating and integrating the multiple semantic features associated with objects; (ii) identification of natural kinds of objects is more reliant on integrating multiple features; (iii) natural kinds of object therefore result in increased anterior temporal activation relative to man-made objects; but (iv) these differential effects of category may only emerge during some tasks. In other tasks, such as the within-category semantic differentiation task used by Devlin et al. (2000), man-made items elicit the same degree of activation as natural kinds.
Conclusion and implications for theoretical models underlying category-specific deficits
In conclusion, previous studies that have compared activation elicited by semantic task and stimulus category have failed to find any overlap between functional knowledge and tool processing. Our study therefore focused on the motor knowledge required for action retrieval. We found that LPMT was activated by (i) the retrieval of action knowledge regardless of stimulus category; and (ii) tools, relative to fruit, even when the task did not require action retrieval. These results are consistent with the proposal that categorizing an object as a tool involves the retrieval of action/motor knowledge, a process that may be mediated by LPMT, while the SMA may have a role in imagining and planning the motor response. LPMT is part of a fronto-parietal system associated with visually presented hand movements (Hërmsdorfer et al., 2001), and the SMA is an established part of the motor system (Dum and Strick, 1991). Thus, the fronto-parietal lesions that have caused deficits with tools more than living kinds may have disrupted connections between LPMT, the SMA and other sensory–motor areas involved in the ideation and generation of action.
The effects for fruits were less clear cut. We did not identify areas specific to perceptual knowledge, and the only fruit-specific effect was in the left medial anterior temporal cortex. We report this finding, although it did not reach a corrected level of significance, because it is consistent with the neuropsychological (Gainotti et al., 2000) and some of the previous neuroimaging literature (Mummery et al., 1996; Moore and Price, 1999; Devlin et al., 2002b). We propose that activation in the left anterior temporal cortex is enhanced for fruit relative to tools because, during some tasks, fruits may increase the demands on ‘semantic integration’. Further studies are required to demonstrate that fruit-specific activation is functionally equivalent to activation associated with semantic integration.
Overall, the data are partially consistent with Warrington and McCarthy’s (1987) model, in that a neural system common to action knowledge and tool discrimination was demonstrated, but the role of perceptual/sensory knowledge in object identification was not substantiated. Instead, we suggest that a hybrid model that distinguishes between identification based on motor properties and identification through conjunctions of semantic features provides a better fit to the category-specific functional neuroimaging data and goes further in explaining the category-specific semantic deficits observed in patients.
|
Appendix |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionAppendixReferences |
|---|
|
|
|
|
|
|
|
|
References |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionAppendixReferences |
|---|
Bottini G, Cocoran R, Sterzi R, Paulesu E, Schenone P, Scarpa P, et al. The role of the right hemisphere in the interpretation of figurative aspects of language. A positron emission tomography activation study. Brain 1994; 117: 1241–53.
Buxbaum LJ, Saffran EM. Knowing ‘how’ vs. ‘what for’: a new dissociation. Brain Lang 1998; 65: 77–86.
Buxbaum LJ, Veramonti T, Schwartz MF. Function and manipulation tool knowledge in apraxia: knowing ‘what for’ but not ‘how’. Neurocase 2000; 6: 83–97.[Web of Science]
Cappa SF, Perani D, Schnur T, Tettamanti M, Fazio F. The effects of semantic category and knowledge type on lexical–semantic access: a PET study. Neuroimage 1998; 8: 350–9.[Web of Science][Medline]
Caramazza A, Shelton JR. Domain-specific knowledge systems in the brain: the animate–inanimate distinction. J Cogn Neurosci 1998; 10: 1–34.[Web of Science][Medline]
Castelli F, Happe F, Frith U, Frith C. Movement and mind: a functional imaging study of perception and interpretation of complex intentional movement patterns. Neuroimage 2000; 12: 314–25.[Web of Science][Medline]
Chao LL, Martin A. Representation of manipulable man-made objects in the dorsal stream. Neuroimage 2000; 12: 478–84.[Web of Science][Medline]
Christensen LO, Johannsen P, Sinkjaer T, Petersen N, Pyndt HS, Nielsen JB. Cerebral activation during bicycle movements in man. Exp Brain Res 2000; 135: 66–72.[Web of Science][Medline]
D‘Esposito M, Detre JA, Aguirre GK, Stallcup M, Alsop DC, Tippet LJ, et al. A functional MRI study of mental image generation. Neuropsychologia 1997; 35: 725–30.[Web of Science][Medline]
Damasio H, Grabowski TJ, Tranel D, Hichwa RD, Damasio AR. A neural basis for lexical retrieval. Nature 1996; 380: 499–505.[Medline]
Decety J, Perani D, Jeannerod M, Bettinardi V, Tadary B, Woods R, et al. Mapping motor representations with positron emission tomography. Nature 1994; 371: 600–2.[Medline]
Decety J, Grezes J, Costes N, Perani D, Jeannerod M, Procyk E, et al. Brain activity during observation of actions. Influence of action content and subject’s strategy. Brain 1997; 120: 1763–77.
Desmond JE, Fiez JA. Neuroimaging studies of the cerebellum: language, learning and memory. Trends Cogn Sci 1998; 2: 355–62.[Web of Science]
Desmond JE, Gabrieli JD, Glover GH. Dissociation of frontal and cerebellar activity in a cognitive task: evidence for a distinction between selection and search. Neuroimage 1998; 7: 368–76.[Web of Science][Medline]
Devlin JT, Gonnerman LM, Andersen ES, Seidenberg MS. Category-specific semantic deficits in focal and widespread brain damage: a computational account. J Cogn Neurosci 1998; 10: 77–94.[Web of Science][Medline]
Devlin JT, Russell RP, Davis MH, Price CJ, Wilson J, Moss HE, et al. Susceptibility-induced loss of signal: comparing PET and fMRI on a semantic task. Neuroimage 2000; 11: 588–600.
Devlin JT, Russell RP, Davis MH, Price CJ, Wilson J, Moss HE, et al. Is there an anatomical basis for category specificity? Semantic memory studies in PET and fMRI. Neuropsychologia 2002a; 40: 54–75.[Web of Science][Medline]
Devlin JT, Moore CJ, Mummery CJ, Gorno-Tempini M, Phillips JA, Noppeney U, et al. Anatomic constraints on cognitive theories of category specificity. Neuroimage 2002b; 15: 675–85.[Web of Science][Medline]
Dum RP, Strick PL. The origin of corticospinal projections from the premotor areas in the frontal lobe. J Neurosci 1991; 11: 667–89.[Abstract]
Fiez JA, Raichle ME, Balota DA, Tallal P, Petersen SE. PET activation of posterior temporal regions during auditory word presentation and verb generation. Cereb Cortex 1996; 6: 1–10.[Medline]
Fink GR, Frackowiak RS, Pietrzyk U, Passingham RE. Multiple nonprimary motor areas in the human cortex. J Neurophysiol 1997; 77: 2164–74.
Fletcher PC, Happe F, Frith U, Baker SC, Dolan RJ, Frackowiak RSJ, et al. Other minds in the brain: a functional imaging study of ‘Theory of mind’ in story comprehension. Cognition 1995; 57: 109–28.[Web of Science][Medline]
Forde EM, Francis D, Riddoch MJ, Rumiati RI, Humphreys GW. On the links between visual knowledge and naming: a single case study of a patient with a category-specific impairment for living things. Cogn Neuropsychol 1997; 14: 403–58.[Web of Science]
Friston KJ, Ashburner J, Frith CD, Poline J-B, Heather JD, Frackowiak RS. Spatial registration and normalization of images. Hum Brain Mapp 1995a; 3: 165–89.[Web of Science]
Friston KJ, Holmes A, Worsley KJ, Poline JB, Frith CD, Frackowiak R. Statistical parametric mapps in functional imaging: a general linear approach. Hum Brain Mapp 1995b; 2: 189–210.[Medline]
Gabrieli JD, Poldrack RA, Desmond JE. The role of left prefrontal cortex in language and memory. [Review]. Proc Natl Acad Sci USA 1998; 95: 906–13.
Gainotti G. What the locus of brain lesion tells us about the nature of the cognitive defect underlying category-specific disorders: a review. [Review]. Cortex 2000; 36: 539–59.[Web of Science][Medline]
Grabowski TJ, Damasio H, Damasio AR. Premotor and prefrontal correlates of category-related lexical retrieval. Neuroimage 1998; 7: 232–43.[Web of Science][Medline]
Grafton ST, Fadiga L, Arbib MA, Rizzolatti G. Premotor cortex activation during observation and naming of familiar tools. Neuroimage 1997; 6: 241–6.
Grézes J, Decety J. Functional anatomy of execution, mental simulation, observation, and verb generation of actions: a meta-analysis. Hum Brain Mapp 2001; 12: 1–19.[Web of Science][Medline]
Grézes J, Costes N, Decety J. Top-down effect of strategy on the perception of human biological motion: a PET investigation. Cogn Neuropsychol 1998; 15: 553–82.[Web of Science]
Grézes J, Costes N, Decety J. The effects of learning and intention on the neural network involved in the perception of meaningless actions. Brain 1999; 122: 1875–87.
Hécaen H, Ajuriaguerra J de. Agnosie visuelle pour les objets inanimés par lésion unilatéral gauche. Rev Neurol (Paris) 1956; 94: 222–33.[Medline]
Hermsdorfer J, Goldenberg G, Wachsmuth C, Conrad B, Ceballos-Baumann AO, Bartenstein P, et al. Cortical correlates of gesture processing: clues to the cerebral mechanisms underlying apraxia during the imitation of meaningless gestures. Neuroimage 2001; 14: 149–61.[Web of Science][Medline]
Humphreys GW, Forde EM. Hierarchies, similarity and interactivity in object recognition: ‘category-specific’ neuro psychological deficits. Behav Brain Sci 2001; 24: 453–509.[Web of Science][Medline]
Klein D, Milner B, Zatorre RJ, Meyer E, Evans AC. The neural substrates underlying word generation: a bilingual functional imaging study. Proc Natl Acad Sci USA 1995; 92: 2899–903.
Kourtzi Z, Kanwisher N. Activation in human MT/MST by static images with implied motion. J Cogn Neurosci 2000; 12: 48–55.[Web of Science][Medline]
Lambon Ralph MA, Howard D, Nightingale G, Ellis AW. Are living and non-living category-specific deficits causally linked to impaired perceptual or associative knowledge? Evidence from a category-specific double dissociation. Neurocase 1998; 4: 311–38.[Web of Science]
Magnié MN, Ferreira CT, Giusiano B, Poncet M. Category specificity in object agnosia: preservation of sensorimotor experiences related to objects. Neuropsychologia 1999; 37: 67–74.[Web of Science][Medline]
Martin A, Haxby JV, Lalonde FM, Wiggs CL, Ungerleider LG. Discrete cortical regions associated with knowledge of color and knowledge of action. Science 1995; 270: 102–5.
Martin A, Wiggs CL, Ungerleider LG, Haxby JV. Neural correlates of category-specific knowledge. Nature 1996; 379: 649–52.[Medline]
McRae K, Cree GS. Factors underlying category-specific semantic deficits. In: Humphreys GW, Forde E, editors. Category-specificity in mind and brain. Hove (UK): Psychology Press. In press 2002
Moore CJ, Price CJ. A functional neuroimaging study of the variables that generate category-specific object processing differences. Brain 1999; 122: 943–62.
Moss HE, Tyler LK, Durrant-Peatfield M, Bunn EM. ‘Two eyes of a see-through’: impaired and intact semantic knowledge in a case of selective deficit for living things. Neurocase 1998; 4: 291–310.[Web of Science]
Mummery CJ, Patterson K, Hodges JR, Wise RJ. Generating ‘tiger’ as an animal name or a word beginning with T: differences in brain activation. Proc R Soc Lond B Biol Sci 1996; 263: 989–95.[Medline]
Mummery CJ, Patterson K, Hodges JR, Price CJ. Functional neuroanatomy of the semantic system: divisible by what? J Cogn Neurosci 1998; 10: 766–77.[Web of Science][Medline]
Nielson JM. Agnosia, apraxia, and aphasia: their value in cerebral localization. 2nd edn. New York: Hoeber; 1946.
Oram MW, Perrett DI. Integration of form and motion in the anterior superior temporal polysensory area (STPa) of the macaque monkey. J Neurophysiol 1996; 76: 109–29.
Peigneux P, Salmon E, van der Linden M, Garraux G, Aerts J, Delfiore G, et al. The role of lateral occipitotemporal junction and area MT/V5 in the visual analysis of upper-limb postures. Neuroimage 2000; 11: 644–55.[Web of Science][Medline]
Perani D, Cappa SF, Bettinardi V, Bressi S, Gorno-Tempini M, Matarrese M, et al. Different neural systems for the recognition of animals and man-made tools. Neuroreport 1995; 6: 1637–41.[Web of Science][Medline]
Perani D, Schnur T, Tettamanti M, Gorno-Tempini ML, Cappa SF, Fazio F. Word and picture matching: a PET study of semantic category effects. Neuropsychologia 1999; 37: 293–306.[Web of Science][Medline]
Perrett DI, Harries MH, Bevan R, Thomas S, Benson PJ, Mistlin AJ, et al. Frameworks of analysis for the neural representation of animate objects and actions. [Review]. J Exp Biol 1989; 146: 87–113.
Petersen SE, Fox PT, Posner MI, Mintun M, Raichle ME. Positron emission tomographic studies of the cortical anatomy of single-word processing. Nature 1988; 331: 585–9.[Medline]
Petersen SE, Fox PT, Posner MI, Mintun M, Raichle ME. Positron emission tomographic studies of the cortical anatomy of single word processing. J Cogn Neurosci 1989; 1: 153–70.
Petit L, Courtney SM, Ungerleider LG, Haxby JV. Sustained activity in the medial wall during working memory delays. J Neurosci 1998; 18: 9429–37.
Phillips JA, Humphreys GW, Noppeney U, Price CJ. The neural substrates of action retrieval: an examination of semantic and visual routes to action. Visual Cognit 2002; 9: 435–46.
Picard N, Strick PL. Motor areas of the medial wall: a review of their location and functional activation. [Review]. Cereb Cortex 1996; 6: 342–53.
Price CJ, Friston KJ. Cognitive conjunction: a new approach to brain activation experiments. Neuroimage 1997; 5: 261–70.[Web of Science][Medline]
Puce A, Allison T, Bentin S, Gore JC, McCarthy G. Temporal motor cortex activation in humans viewing eye and mouth movements. [Review]. J Neurosci 1998; 18: 2188–99.
Riddoch MJ, Humphreys GW. Visual object processing in a case of optic aphasia: a case of semantic agnosia. Cogn Neuropsychol 1987; 4: 131–85.[Web of Science]
Rizzolatti G, Fadiga L, Matelli M, Bettinardi V, Paulesu E, Perani D, et al. Localization of group representations in humans by PET: 1. Observation versus execution. Exp Brain Res 1996; 111: 246–52.[Web of Science][Medline]
Sakai K, Hikosaka O, Miyauchi S, Sasaki Y, Fujimaki N, Putz B. Presupplementary motor area activation during sequence learning reflects visuo-motor association. J Neurosci 1999; 19: RC1. Available from: htp: //www.jneurosci.org
Schmahmann JD, Doyon J, McDonald D, Holmes C, Lavoie K, Hurwitz AS, et al. Three dimensional MRI atlas of the human cerebellum in proportional stereotaxic space. Neuroimage 1999; 10: 233–60.[Web of Science][Medline]
Seltzer B, Pandya DN. Afferent cortical connections and architectonics of the superior temporal sulcus and surrounding cortex in the rhesus monkey. Brain Res 1978; 149: 1–24.[Web of Science][Medline]
Seltzer B, Pandya DN. Parietal, temporal, and occipital projections to cortex of the superior temporal sulcus in the rhesus monkey: a retrograde tracer study. J Comp Neurol 1994; 343: 445–63.[Web of Science][Medline]
Senior C, Barnes J, Giampietro V, Simmons A, Bullmore ET, Brammer M, et al. The functional neuroanatomy of implicit-motion perception or representational momentum. Curr Biol 2000; 10: 16–22.[Web of Science][Medline]
Sirigu A, Duhamel J-R, Poncet M. The role of sensorimotor experience in object recognition: a case of multimodal agnosia. Brain 1991; 114: 2555–73.
Snodgrass JG, Vanderwart M. A standardized set of 260 pictures: norms for name agreement, image agreement, familiarity, and visual complexity. J Exp Psychol [Hum Learn] 1980; 6: 174–215.[Medline]
Stephan KM, Fink GR, Passingham RE, Silbersweig D, Ceballos-Baumann AO, Frith CD, et al. Functional anatomy of the mental representation of upper extremity movements in healthy subjects. J Neurophysiol 1995; 73: 373–86.
Tanji J, Shima K. Role for supplementary motor area cells in planning several movements ahead. Nature 1994; 371: 413–6.[Medline]
Thompson-Schill SL, Aguirre GK, D’Esposito M, Farah MJ. A neural basis for category and modality specificity of semantic knowledge. Neuropsychologia 1999; 37: 671–6.[Web of Science][Medline]
Tippett LJ, Gloser G, Farah MJ. A category-specific naming impairment after temporal lobectomy. Neuropsychologia 1996; 34: 139–46.[Web of Science][Medline]
Tranel D, Damasio H, Damasio AR. A neural basis for the retrieval of conceptual knowledge. Neuropsychologia 1997; 35: 1319–27.[Web of Science][Medline]
Vandenberghe R, Price C, Wise R, Josephs O, Frackowiak RS. Functional anatomy of a common semantic system for words and pictures. Nature 1996; 383: 254–6.[Medline]
Warburton E, Wise RJ, Price CJ, Weiller C, Hadar U, Ramsay S, et al. Noun and verb retrieval by normal subjects. Studies with PET. [Review]. Brain 1996; 119: 159–79.
Warrington E, McCarthy R. Categories of knowledge: further fractionation and an attempted integration. Brain 1987; 11: 1273–96.
Warrington EK, Shallice T. Category specific semantic impairments. Brain 1984; 107: 829–53.
Watson RT, Fleet WS, Gonzalez-Rothi L, Heilman KM. Apraxia and the supplementary motor area. Arch Neurol 1986; 43: 787–92.
Watson JD, Myers R, Frackowiak RS, Hajnal JV, Woods RP, Mazziotta JC, et al. Area V5 of the human brain: evidence from a combined study using positron emission tomography and magnetic resonance imaging. Cereb Cortex 1993; 3: 79–94.
Yamadori A, Albert M. Word category aphasia. Cortex 1973; 9: 112–25.[Medline]
Zeki S, Watson JD, Frackowiak RS. Going beyond the information given: the relation of illusory visual motion to brain activity. Proc R Soc Lond B Biol Sci 1993; 252: 215–22.[Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  J. R. Binder, R. H. Desai, W. W. Graves, and L. L. Conant Where Is the Semantic System? A Critical Review and Meta-Analysis of 120 Functional Neuroimaging Studies Cereb Cortex, March 27, 2009; (2009) bhp055v1. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Hocking and C. J. Price The Role of the Posterior Superior Temporal Sulcus in Audiovisual Processing Cereb Cortex, October 1, 2008; 18(10): 2439 - 2449. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. Canessa, F. Borgo, S. F. Cappa, D. Perani, A. Falini, G. Buccino, M. Tettamanti, and T. Shallice The Different Neural Correlates of Action and Functional Knowledge in Semantic Memory: An fMRI Study Cereb Cortex, April 1, 2008; 18(4): 740 - 751. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. J. H. Ebisch, C. Babiloni, C. Del Gratta, A. Ferretti, M. G. Perrucci, M. Caulo, M. M. Sitskoorn, and G. L. Romani Human Neural Systems for Conceptual Knowledge of Proper Object Use: A Functional Magnetic Resonance Imaging Study Cereb Cortex, November 1, 2007; 17(11): 2744 - 2751. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Kiefer, S. Schuch, W. Schenck, and K. Fiedler Mood States Modulate Activity in Semantic Brain Areas during Emotional Word Encoding Cereb Cortex, July 1, 2007; 17(7): 1516 - 1530. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. W. Lewis Cortical Networks Related to Human Use of Tools Neuroscientist, June 1, 2006; 12(3): 211 - 231. [Abstract] [PDF]
|
|
|
|
 |
|
 J. W. Lewis, J. A. Brefczynski, R. E. Phinney, J. J. Janik, and E. A. DeYoe Distinct Cortical Pathways for Processing Tool versus Animal Sounds J. Neurosci., May 25, 2005; 25(21): 5148 - 5158. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. W. Lewis, F. L. Wightman, J. A. Brefczynski, R. E. Phinney, J. R. Binder, and E. A. DeYoe Human Brain Regions Involved in Recognizing Environmental Sounds Cereb Cortex, September 1, 2004; 14(9): 1008 - 1021. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 U. Noppeney, K. J. Friston, and C. J. Price Effects of visual deprivation on the organization of the semantic system Brain, July 1, 2003; 126(7): 1620 - 1627. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||